Regularized Distribution Matching Distillation for One-step Unpaired Image-to-Image Translation
Fuente:
arXiv
Salvato in:
| Autori principali: | , , |
|---|---|
| Natura: | Preprint |
| Pubblicazione: |
2024
|
| Soggetti: | |
| Accesso online: | |
| Tags: |
Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
|
| _version_ | 1866917700292313088 |
|---|---|
| author | Rakitin, Denis Shchekotov, Ivan Vetrov, Dmitry |
| author_facet | Rakitin, Denis Shchekotov, Ivan Vetrov, Dmitry |
| contents | Diffusion distillation methods aim to compress the diffusion models into efficient one-step generators while trying to preserve quality. Among them, Distribution Matching Distillation (DMD) offers a suitable framework for training general-form one-step generators, applicable beyond unconditional generation. In this work, we introduce its modification, called Regularized Distribution Matching Distillation, applicable to unpaired image-to-image (I2I) problems. We demonstrate its empirical performance in application to several translation tasks, including 2D examples and I2I between different image datasets, where it performs on par or better than multi-step diffusion baselines. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2406_14762 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Regularized Distribution Matching Distillation for One-step Unpaired Image-to-Image Translation Rakitin, Denis Shchekotov, Ivan Vetrov, Dmitry Computer Vision and Pattern Recognition Machine Learning Diffusion distillation methods aim to compress the diffusion models into efficient one-step generators while trying to preserve quality. Among them, Distribution Matching Distillation (DMD) offers a suitable framework for training general-form one-step generators, applicable beyond unconditional generation. In this work, we introduce its modification, called Regularized Distribution Matching Distillation, applicable to unpaired image-to-image (I2I) problems. We demonstrate its empirical performance in application to several translation tasks, including 2D examples and I2I between different image datasets, where it performs on par or better than multi-step diffusion baselines. |
| title | Regularized Distribution Matching Distillation for One-step Unpaired Image-to-Image Translation |
| topic | Computer Vision and Pattern Recognition Machine Learning |
| url | https://arxiv.org/abs/2406.14762 |